ABSTRACT GABAergic neurons are the main inhibitory neurons in the adult nervous system. There are diverse cortical and subcortical GABAergic interneurons and projection neurons with different physiological properties, connectivity, and functions. However, few studies have been able to investigate the functions of specific subclasses by perturbing GABA release selectively. We propose an approach to target GABA synthesis in subclasses of GABAergic neurons. GAD activity requires the cofactor pyridoxal-5-phosphate (PLP). PLP deficiency can be achieved by knocking out the rate limiting enzyme in PLP synthesis: pyridoxine-5-phosphate oxidase (PNPO). Moreover, PLP deficiency can be rescued by systemic PLP administration. We therefore propose to generate cell type specific PNPO deficiency mouse models. We will first generate models with GABA deficiency in well-studied but highly distinct cell types to test the feasibility of our approach: 1) the parvalbumin (PV)-positive GABAergic neurons; 2) the VIP-positive GABAergic neurons; 3) the striatal dopamine D1 receptor expressing neurons; and 4) the striatal dopamine D2 receptor expressing neurons. Mice will be assessed for general health, feeding, drinking, sleep, and potential seizures. In addition, working memory, reference memory, reversal learning, motor coordination and motor learning will be examined. Brain morphology and electrophysiology will be assessed. Levels of GABA, other PNPO-dependent neurotransmitters, and their metabolites will be measured. The main goal of this R21 application is to test the feasibility of such an approach. These models are potentially significant research tools for the research community and may have a broad impact given the diverse distribution and functions of different GABAergic projection neurons and interneurons.